The proposed work will develop a general framework for the ecological and evolutionary dynamics of "disease communities". Although the field of epidemiology has a distinguished history with notable successes, the potential for interaction between unrelated infections has not received much attention. A mechanism for interaction between antigenically distinct infections is proposed: following an acute infection, individuals are temporarily unavailable to contract other diseases (primarily because quarantining during convalescence). Hence, the number of potential "hosts" available for each pathogen is affected by the outbreak dynamics of other infections. If an infection is associated with substantial mortality, then potential hosts become permanently removed and the interaction takes the form of competition between diseases. This mechanism leads to what is called "disease interference". This proposal aims to develop this conceptual framework, answering a number of fundamental ecological and evolutionary questions. Do all infections interact with all other infections? The answer to this is clearly no. Intuitively, it would be expected that the strength of this negative interaction between infections would depend on the degree similarity in hosts infected. For human infections, this would be determined by the amount of overlap between the distributions of the host age at infection. Does disease interference affect evolutionary dynamics? A central hypothesis of this work is that interference is likely to select for increased disease virulence. Can we use interference effects in systems where antigenic polymorphism is well established (e.g. Dengue)? The interference mechanism provides a null model for the study of infections with multiple strains. Is this work likely to have any important public health implications? Preliminary work suggests interference effects may be beneficially used to eradicate infections using fewer vaccine units than using conventional estimates. This proposal also aims to construct statistical tools whereby the signature of interference may be confidently detected from data. Due to their excellent spatio-temporal data sets, much of this work will focus on childhood infections (e.g. as measles, pertussis, chickenpox and rubella), though the proposed mechanism is quite general. The analytical tools developed will be applied to numerous long-term disease records to explore the ubiquity of the interference phenomenon.